{"id":"W4400485045","doi":"10.1145/3664646.3664772","title":"Chain of Targeted Verification Questions to Improve the Reliability of Code Generated by LLMs","year":2024,"lang":"en","type":"article","venue":"","topic":"Software Testing and Debugging Techniques","field":"Computer Science","cited_by":4,"is_retracted":false,"has_abstract":true,"ca_institutions":"Polytechnique Montréal","funders":"","keywords":"Reliability (semiconductor); Computer science; Code (set theory); Chain (unit); Reliability engineering; Risk analysis (engineering); Programming language; Business; Engineering","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0006046114,0.00006488799,0.00008854782,0.00005258465,0.00003889339,0.00003806259,0.0004194296,0.00003741638,0.000005016123],"category_scores_gemma":[0.0005261008,0.00004271772,0.0000317914,0.0005896968,0.00004648231,0.00008247874,0.00008286112,0.00006875914,0.000004993585],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00002459203,"about_ca_system_score_gemma":0.00006659079,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0005383131,"about_ca_topic_score_gemma":0.000004198278,"domain_scores_codex":[0.999248,0.00006623345,0.000219758,0.0002384383,0.0001351593,0.00009238466],"domain_scores_gemma":[0.9988885,0.0002547277,0.00004255552,0.0006129512,0.0001697694,0.00003151515],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00001070224,0.0002241941,0.001000914,0.0001428068,0.00002893023,0.000001266951,0.001969436,0.0008529021,0.4939511,0.0618696,0.3640781,0.07586999],"study_design_scores_gemma":[0.00004885243,0.0002035301,0.001427326,0.0000686534,0.000008405085,0.000002093201,0.000006016856,0.5264503,0.4446366,0.02380476,0.003205279,0.0001381993],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.03238045,0.0001614126,0.9609082,0.00231722,0.0001747356,0.0002066195,0.00002283579,0.003636967,0.0001915846],"genre_scores_gemma":[0.8263441,0.000004064595,0.1732891,0.0001035581,0.00001078586,0.00003205228,0.000004415073,0.000004642237,0.0002072838],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.7939636,"threshold_uncertainty_score":0.1741978,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01257638812373305,"score_gpt":0.2681475290310201,"score_spread":0.255571140907287,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}